def grid_search_best_model_timeseries_ar(df, grid, cv):
best_param = None
best_score = np.infty
tsp = TimeSeriesSplit(n_splits=cv)
for param in grid.get('lags'):
scores = []
for train_ind, test_ind in tsp.split(df):
train_data = df.iloc[train_ind]
test_data = df.iloc[test_ind]
try:
#print(train_data, test_data)
estimator = ar_model.AutoReg(train_data, lags=param)
res = estimator.fit()
#print(res.params)
#get out of sample predictions with test data start and end
pred = estimator.predict(res.params, test_data.index[0],
test_data.index[-1])
#print(pred)
y_pred = pred.values.reshape(-1)
y_test = test_data.values.reshape(-1)
score = math.sqrt(metrics.mean_squared_error(y_test, y_pred))
scores.append(score)
except:
pass
#print(scores)
if len(scores) > 0 and np.mean(scores) < best_score:
best_score = np.mean(scores)
best_param = param
if best_param is not None:
estimator = ar_model.AutoReg(df, lags=best_param)
res = estimator.fit()
print("best parameters:" + str(best_param))
print("validation rmse:" + str(best_score))
#get insample predictions with start and end indices
predictions = estimator.predict(res.params,
start=0,
end=df.shape[0] - 1)
y_pred = predictions.values.reshape(-1)
y_train = df.values.reshape(-1)[best_param:]
train_rmse = math.sqrt(metrics.mean_squared_error(y_train, y_pred))
print("train rmse:" + str(train_rmse))
return estimator, res
else:
return None, None
#read our petrol data in
series = read_csv('petrol_prices.csv',
header=0,
index_col=0,
parse_dates=True,
squeeze=True)
# split dataset
X = difference(series.values)
size = int(len(X) * 0.66)
train, test = X[0:size], X[size:]
# train autoregression
window = 10
model = ar_model.AutoReg(train, lags=6)
model_fit = model.fit()
coef = model_fit.params
# walk forward over time steps in test
history = [train[i] for i in range(len(train))]
predictions = list()
for t in range(len(test)):
yhat = predict(coef, history)
obs = test[t]
predictions.append(yhat)
history.append(obs)
rmse = sqrt(mean_squared_error(test, predictions))
#the mean sq. error
print('Test RMSE: %.3f' % rmse)
prepped_dataf['New Cases'] = prepped_dataf['Log Number Cases Delta']
prepped_dataf = prepped_dataf[prepped_dataf['New Cases'].notna()]
prepped_dataf = prepped_dataf.replace([np.inf, -np.inf], np.nan)
prepped_dataf = prepped_dataf[prepped_dataf['New Cases'].notna()]
logged_vals = prepped_dataf['Log Number Cases']
for col in prepped_dataf.keys():
if re.match(r'Log.*', col):
print(re.match(r'Log.*', col))
prepped_dataf = prepped_dataf.drop(col, axis=1)
with pandas.option_context('display.max_rows', None, 'display.max_columns',
None): # more options can be specified also
print(prepped_dataf)
#prepped_dataf.drop(prepped_dataf.tail(3).index,inplace=True) # drop last n rows
model = ar_model.AutoReg(prepped_dataf, lags=2).fit()
print(model.summary())
target_len = len(prepped_dataf['New Cases']) - 1
prediction = model.predict(start=205, end=205 + target_len)
predictions = [pred for pred in prediction]
cases_log_diff_rev = []
for v in predictions:
index = predictions.index(v)
v_0 = logged_vals[index]
v_0 += v
cases_log_diff_rev.append(v_0)
print(np.exp(cases_log_diff_rev))
start_date = prepped_dataf.index[0]
start_date = start_date + datetime.timedelta(days=1)
date_list = [
energy_test = energy[energy.index >= test_start_dt]
scaler = MinMaxScaler()
energy_train['load_scaled'] = scaler.fit_transform(energy_train['load'])
energy_train.head(10)
#plot both
sns.distplot(energy_train['load'])
sns.distplot(energy_train['load_scaled'])
energy_train1 = energy_train.copy()
energy_train1 = energy_train1.drop('load', axis=1)
energy_train1.index.freq = 'H'
#build model
estimator = ar_model.AutoReg(energy_train1, lags=5)
res = estimator.fit()
print(res.params)
energy_validation['load_scaled'] = scaler.fit_transform(energy_validation)
energy_validation1 = energy_validation.copy()
energy_validation1 = energy_validation1.drop('load', axis=1)
energy_validation1.index.freq = 'H'
#valdiation error
pred = estimator.predict(res.params, energy_validation1.index[0],
energy_validation1.index[-1])
print(pred)
y_pred = pred.values.reshape(-1)
y_test = energy_validation1.values.reshape(-1)
score = math.sqrt(metrics.mean_squared_error(y_test, y_pred))
import os
from sklearn import metrics
from sklearn.model_selection import TimeSeriesSplit
from statsmodels.tsa import ar_model
path = 'F:/'
df = pd.read_csv(os.path.join(path, 'uk-deaths-from-bronchitis-emphys.csv'))
df.info()
df.columns = ['timestamp', 'y']
df.index = pd.to_datetime(df['timestamp'], format='%Y-%m').copy()
df.index.freq = 'MS'
df.drop('timestamp', axis=1, inplace=True)
#build model
estimator = ar_model.AutoReg(df, lags=5)
res = estimator.fit()
print(res.params)
print(res.model)
print(res.summary())
#using model
predictions = estimator.predict(res.params, start=0, end=df.shape[0] - 1)
print(predictions)
y_pred = predictions.values.reshape(-1)
y_train = df.values.reshape(-1)[5:]
train_rmse = math.sqrt(metrics.mean_squared_error(y_train, y_pred))
print(train_rmse)
#evaluate model
tsp = TimeSeriesSplit(n_splits=3)
def train(
data: np.ndarray,
used_model: str = "autoreg",
p: int = 5,
d: int = 1,
q: int = 0,
cov_type="nonrobust",
method="cmle",
trend="nc",
solver="lbfgs",
maxlag=13,
# SARIMAX args
seasonal=(0, 0, 0, 0),
) -> Any:
"""Autoregressive model from statsmodels library. Only univariate data.
Args:
data (np.ndarray): Time series data.
used_model (str, optional): Used model. Defaults to "autoreg".
p (int, optional): Order of ARIMA model (1st - proportional). Check statsmodels docs for more. Defaults to 5.
d (int, optional): Order of ARIMA model. Defaults to 1.
q (int, optional): Order of ARIMA model. Defaults to 0.
cov_type: Parameters of model call or fit function of particular model. Check statsmodels docs for more.
Defaults to 'nonrobust'.
method: Parameters of model call or fit function of particular model. Check statsmodels docs for more.
Defaults to 'cmle'.
trend: Parameters of model call or fit function of particular model. Check statsmodels docs for more.
Defaults to 'nc'.
solver: Parameters of model call or fit function of particular model. Check statsmodels docs for more.
Defaults to 'lbfgs'.
maxlag: Parameters of model call or fit function of particular model. Check statsmodels docs for more.
Defaults to 13.
seasonal: Parameters of model call or fit function of particular model. Check statsmodels docs for more.
Defaults to (0, 0, 0, 0).
Returns:
statsmodels.model: Trained model.
"""
import statsmodels.tsa.api as sm
from statsmodels.tsa.statespace.sarimax import SARIMAX
from statsmodels.tsa.arima.model import ARIMA
from statsmodels.tsa import ar_model
used_model = used_model.lower()
if used_model == "ar":
model = sm.AR(data)
fitted_model = model.fit(method=method, trend=trend, solver=solver, disp=0)
elif used_model == "arima":
order = (p, d, q)
model = ARIMA(data, order=order)
fitted_model = model.fit()
elif used_model == "sarimax":
order = (p, d, q)
model = SARIMAX(data, order=order, seasonal_order=seasonal)
fitted_model = model.fit(method=method, trend=trend, solver=solver, disp=0)
elif used_model == "autoreg":
auto = ar_model.ar_select_order(data, maxlag=maxlag)
model = ar_model.AutoReg(
data,
lags=auto.ar_lags,
trend=auto.trend,
seasonal=auto.seasonal,
period=auto.period,
)
fitted_model = model.fit(cov_type=cov_type)
else:
raise ValueError(
f"Used model has to be one of ['ar', 'arima', 'sarimax', 'autoreg']. You configured: {used_model}"
)
setattr(fitted_model, "my_name", used_model)
setattr(fitted_model, "data_length", len(data))
return fitted_model